TransPose is a human pose estimation model based on a CNN feature extractor, a Transformer Encoder, and a prediction head. Given an image, the attention layers built in Transformer can efficiently capture long-range spatial relationships between keypoints and explain what dependencies the predicted keypoints locations highly rely on.

[arxiv 2012.14214] [paper] [demo-notebook]

TransPose: Keypoint Localization via Transformer, Sen Yang, Zhibin Quan, Mu Nie, Wankou Yang, ICCV 2021

We choose two types of CNNs as the backbone candidates: ResNet and HRNet. The derived convolutional blocks are ResNet-Small, HRNet-Small-W32, and HRNet-Small-W48.

Try out the Web Demo:

You can directly load TransPose-R-A4 or TransPose-H-A4 models with pretrained weights on COCO train2017 dataset from Torch Hub, simply by:

import torch

tpr = torch.hub.load('yangsenius/TransPose:main', 'tpr_a4_256x192', pretrained=True)
tph = torch.hub.load('yangsenius/TransPose:main', 'tph_a4_256x192', pretrained=True)

Note:

• we computed the average FPS* of testing 100 samples from coco val dataset (with batchsize=1) on a single NVIDIA 2080Ti GPU. The FPS may fluctuate up and down at different tests.
• We trained our different models on different hardware platforms: 1 x RTX2080Ti GPUs (TP-R-A4), 4 x TiTan XP GPUs (TP-H-S, TP-H-A4), and 4 x Tesla P40 GPUs (TP-H-A6).

Jupyter Notebook Demo

Given an input image, a pretrained TransPose model, and the predicted locations, we can visualize the spatial dependencies of the predicted locations with threshold for the attention scores.

TransPose-R-A4 with threshold=0.00

TransPose-R-A4 with threshold=0.01

TransPose-H-A4 with threshold=0.00

TransPose-H-A4 with threshold=0.00075

1. Clone this repository, and we'll call the directory that you cloned as ${POSE_ROOT}

   git clone https://github.com/yangsenius/TransPose.git

2. Install PyTorch>=1.6 and torchvision>=0.7 from the PyTorch official website

3. Install package dependencies. Make sure the python environment >=3.7

   pip install -r requirements.txt

4. Make output (training models and files) and log (tensorboard log) directories under ${POSE_ROOT} & Make libs

   mkdir output log
   cd ${POSE_ROOT}/lib
   make

5. Download pretrained models from the releases of this repo to the specified directory

   ${POSE_ROOT}
    `-- models
        `-- pytorch
            |-- imagenet
            |   |-- hrnet_w32-36af842e.pth
            |   |-- hrnet_w48-8ef0771d.pth
            |   |-- resnet50-19c8e357.pth
            |-- transpose_coco
            |   |-- tp_r_256x192_enc3_d256_h1024_mh8.pth
            |   |-- tp_r_256x192_enc4_d256_h1024_mh8.pth
            |   |-- tp_h_32_256x192_enc4_d64_h128_mh1.pth
            |   |-- tp_h_48_256x192_enc4_d96_h192_mh1.pth
            |   |-- tp_h_48_256x192_enc6_d96_h192_mh1.pth    

We follow the steps of HRNet to prepare the COCO train/val/test dataset and the annotations. The detected person results are downloaded from OneDrive or GoogleDrive. Please download or link them to ${POSE_ROOT}/data/coco/, and make them look like this:

${POSE_ROOT}/data/coco/
|-- annotations
|   |-- person_keypoints_train2017.json
|   `-- person_keypoints_val2017.json
|-- person_detection_results
|   |-- COCO_val2017_detections_AP_H_56_person.json
|   `-- COCO_test-dev2017_detections_AP_H_609_person.json
`-- images
	|-- train2017
	|   |-- 000000000009.jpg
	|   |-- ... 
	`-- val2017
		|-- 000000000139.jpg
		|-- ... 

python tools/test.py --cfg experiments/coco/transpose_r/TP_R_256x192_d256_h1024_enc4_mh8.yaml TEST.USE_GT_BBOX True

python tools/train.py --cfg experiments/coco/transpose_r/TP_R_256x192_d256_h1024_enc4_mh8.yaml

Great thanks for these papers and their open-source codes：HRNet, DETR, DarkPose

This repository is released under the MIT LICENSE.

If you find this repository useful please give it a star 🌟 or consider citing our work:

@inproceedings{yang2021transpose,
  title={TransPose: Keypoint Localization via Transformer},
  author={Yang, Sen and Quan, Zhibin and Nie, Mu and Yang, Wankou},
  booktitle={IEEE/CVF International Conference on Computer Vision (ICCV)},
  year={2021}
}